In old-growth forests, seasonal variation in soil respiration has been well known but the spatial variability, especially between gap and canopy areas, has been reported unclearly. The previous studies of soil respiration using soda lime method in an old-growth beech-oak forest at Ohshirakawa showed both seasonal and spatial variations. The effect of soil temperature and soil water content on seasonal soil respiration was not clear. The spatial variation of soil respiration which was higher in canopy areas than in gap areas, was likely influenced by the root biomass which was also higher in canopy areas. This special spatial soil respiration lead to the new objective which is to upscale estimation of soil respiration in a 1-ha study. In the present study, soil respiration was measured using an infrared gas analysis connected to automatic open-close chamber system in gap and canopy areas. The relationships between soil respiration, soil temperature and soil water content were compared for the best fitted model in gap, canopy and all areas. Soil respiration at whole study site was estimated based on information of the spatial gradients of soil respiration and the area of crown closure.